GestureMap: Supporting Visual Analytics and Quantitative Analysis of Motion Elicitation Data by Learning 2D Embeddings
This work addresses the problem of analyzing complex gesture data for researchers in human-computer interaction, though it is incremental as it builds on existing methods like VAEs and DTW for a specific domain.
The paper tackles the challenge of analyzing gesture elicitation data by introducing GestureMap, a visual analytics tool that uses a Variational Autoencoder to embed 3D gesture skeletons into a 2D map, enabling exploration and quantitative analysis through features like average gesture computation and clustering. The result is a tool that facilitates visual understanding of gesture spaces and opens new research directions, as validated by expert evaluations and analysis of published data.
This paper presents GestureMap, a visual analytics tool for gesture elicitation which directly visualises the space of gestures. Concretely, a Variational Autoencoder embeds gestures recorded as 3D skeletons on an interactive 2D map. GestureMap further integrates three computational capabilities to connect exploration to quantitative measures: Leveraging DTW Barycenter Averaging (DBA), we compute average gestures to 1) represent gesture groups at a glance; 2) compute a new consensus measure (variance around average gesture); and 3) cluster gestures with k-means. We evaluate GestureMap and its concepts with eight experts and an in-depth analysis of published data. Our findings show how GestureMap facilitates exploring large datasets and helps researchers to gain a visual understanding of elicited gesture spaces. It further opens new directions, such as comparing elicitations across studies. We discuss implications for elicitation studies and research, and opportunities to extend our approach to additional tasks in gesture elicitation.